Gagne de la cryptomonnaie GRATUITE en 5 clics et aide institut numérique à propager la connaissance universitaire >> CLIQUEZ ICI <<

2.1.1. Perception of English: from suprasegments to segments

Non classé

The acquisition of the L1 phonology occurs in different stages, roughly corresponding to
different ages of the infant. As far as perception capacities are concerned, Kaplan and Kaplan
(1971) notice that the division into distinct stages is not so clear as it can be for production
(cf. 2.1.2. below). Nevertheless, general patterns seem to be recurrent in normally developing
systems. It is increasingly accepted that prosody constitutes the very first contact that the
human being has with language, even intra-uterine. That is why children are believed to start
acquiring prosodic features at a pre-linguistic stage, long before the acquisition of
segmentals. It is only at the end of the first year of life that segments become more important,
especially because they help the child form his/her first words. The infant’s receptive control
over the suprasegmental system emerges before the control over the segmental system of the
language (idem). Even if still in the mother’s belly, the infant can hear the melody that is
created by the prosody of the ambient language. Then, although he/she cannot distinctly
make out individual sounds immediately, rhythm, stress beats, accents, intonation,
contribute to the infant’s first perception of the mother tongue.

Crystal (1970) emphasizes the contribution of prosodic patterns to the development of
grammatical competence. Intonation, particularly, helps the early learner organize utterances
into chunks. Similarly, Gerken (1996) discusses how infants use prosodic information to infer
syntactic structure. It serves to segment the speech stream into major units and to locate the
linguistically relevant ones, so that the child is more and more familiar with the linguistic
system. The experiment(1) by Christophe, Melher and Sebastián-Gallés (2001) illustrates this
prosodic segmentation hypothesis. They found that French newborn infants manage to
discriminate Spanish through phonological phrase boundaries: “Phonological phrase
boundaries often coincide with boundaries of syntactic constituents. Therefore, they may
provide some information as to the syntactic structure of sentences” (idem: 386). Prosodic
features are more and more acknowledged to be acquired very early by infants. Johnson and
Reimers (2010: 94) observe: “Studies have shown that newborns can discriminate languages
with different rhythms […], but not languages belonging to the same linguistic rhythm
group”. Not only does this prove that early learners greatly rely on prosodic features, but it
is all the more interesting as it involves the typological distinction among stress-timed,
syllable-timed, and mora-timed languages. Nonetheless, more investigations of the role of
suprasegmental features at the perception level – whether cross-linguistically or not – need
to be carried out.

As regards the infant’s perception of segmental features, Jusczyk (1992: 20) claims that the
capacity of recognizing voicing contrasts of especially initial stops (e.g. /pa/ vs. /ba/) starts as
early as one month of age: “Infants have the capacity to do some preliminary grouping of
speech sounds into different perceptual categories”. The developing research on the
perceptual skills of infants shows that during the first six months of life, infants can perceive
more phonetic contrasts than merely onset plosives (Johnson & Reimers, 2010: 74). The
reason defended by many researchers is that humans are born with a universal capacity of
categorizing sounds – they have the ability to perceive speech in terms of phonemes (idem:
78). This means that infants can perceive phonetic contrasts that occur in the L1, but also in
any language of the world, as opposed to adults (Werker, 1995: 89). The influence of and
exposure to the mother tongue, however, appear very quickly. Around the age of six months,
speech sound categories that are based on the L1 sounds are formed and develop. While the
newborn has the capacity to discriminate both L1 and L2 contrasts (Carlotti, 2007), this
capacity gradually loses ground, in keeping with the language input to which the child is
exposed in his/her everyday environment. Depending on the ambient language(s), the
influence of the latter will be increasingly important. Yet, the ability to discriminate contrasts
that are not in their environment does not disappear altogether, but children start to perceive
less even at that stage.

Kuhl’s (1991, for example) work is often taken up by other researchers. The author put
forward the Perceptual Magnet Effect hypothesis (abbreviated PME), according to which
there is a strong influence of the L1 phonology on the child’s perception of sounds. As a
consequence, the hypothesis holds that infants create mental representations of the sounds
that they hear, and the most representative sounds of a certain category – called the
“prototypes” – function like perceptual magnets on other sounds of the same category. By
the end of the first year of life, infants thus stop being universal listeners. Their speech
perception performance declines and increasingly matches the L1 sound properties (Werker,
1995: 89). As a result, the child starts responding differently to foreign sounds that he/she
hears. In fact, this loss of sensitivity is prone to debate and arguments; some think that it is
permanent and absolute, and others claim the opposite. That is why it is referred to as the
Maintenance/Loss View (idem: 95).

In the L1 acquisition process, the infant is often said to be a universal listener. He/she can
perceive prosodic and phonetic contrasts of any language in the world. Prosody is the very
first linguistic element with which the infant has a contact. It is only later that he/she starts to
make out individual sounds, which also constitutes the first steps of production.

1 Most of the experiments on the perception of language by infants are conducted thanks to close
examinations of head-orientation responses to natural speech.

Page suivante : 2.1.2. Production of English segments and suprasegments